Add a button to extract the product name from Front Photo using OCR

[teolemon]

Problem

• It's painful to enter the product name
• As a result, many people don't take the time
• Some major apps have taken a scraping/api approach for prefilling (but it's probably not something we want to do, and it takes a lot of time to build)

Proposed solution

• Add a button to extract the product name from Front Photo using OCR

Additional context

• Extraction won't be perfect
• We can use Robotoff or a LLM in the future to clean up the OCR output
• For now, we rely on contributors to cleanup the output, and fix/choose the right language (like in my example where we have a German image selected as English)

Number of products impacted

197000 products (https://world.openfoodfacts.org/state/product-name-to-be-completed)

Time per product

30 seconds

Upside

68 days of round the clock contributions if time reduced to 0 for all of them

Mockups

Part of

• ✏️ Editing (Tracker) #5525